Inequalities in childbirth experiences: A population-based cross-sectional survey in England during the Covid-19 pandemic

Patient and Public Involvement

National maternity surveys are carried out by the National Perinatal Epidemiology Unit (NPEU) as part of an ongoing programme of research. The 2020 survey was undertaken by the Policy Research Unit in Maternal and Neonatal Health and Care (PRU-MNHC) which includes two Parent, Patient and Public Involvement and Engagement (PPPIE) co-leads as part of the core team. The PPPIE co-leads were involved in the design and planning of the 2020 survey from the outset and drew on the vast PPPIE network of the PRU-MNHC to gain additional input to the study. PPPIE input was sought throughout the project and particularly for the design of study materials, selection of questions and measures to be included in the questionnaire, interpretation of survey findings, drafting of lay and infographic summary reports, and planning the dissemination strategy for the survey findings. In addition to the PPPIE strategy which was devised for the 2020 survey, the 2020 survey was based on earlier surveys carried out by the NPEU, which have all relied on extensive PPPIE and user input.

Design and participants

This study analysed data from the 2020 national maternity survey, a large population-based cross-sectional survey of women’s health and maternity care during pregnancy, labour and birth, and the postnatal period. A random sample of 16,050 women was identified by the Office for National Statistics (ONS) from birth registration records. The sample size was based on the response rate to previous surveys taking into account the trend in falling response rates and was selected to ensure we could estimate the prevalence of most outcomes with reasonable precision and have adequate power to compare key outcomes in different groups of women. Women were eligible if they were aged 16 years or over and had given birth in England during May 2020, which was during the first Covid-19 “lockdown” in the UK. Women were six months postpartum at the time of recruitment. The survey was sent out via post in November 2020 and women were able to take part on paper, by telephone (with an interpreter if required), or online. Non-respondents were sent up to two reminders and the survey closed in March 2021. Prior to all mailings, checks on infant deaths were made by ONS and any women whose babies had died were excluded.

Sociodemographic factors

ONS provided sociodemographic data for the whole sample of women to enable comparison of the respondents and non-respondents. These data included maternal age, country of birth, marital status at the time the birth was registered (registration status), and socioeconomic deprivation based on the index of multiple deprivation (IMD), a measure of relative small area-based deprivation across England. In addition to the sociodemographic data supplied by ONS, women who responded to the survey were asked to identify their ethnic group and to indicate the length of time they had lived in the UK if they were born outside the UK. Women were also asked to indicate their age when they completed full-time education.

For the analysis, maternal age was grouped into three categories: younger than 25 years; between 25–34 years; or 35 years and older. Registration status was grouped into three categories: married; joint registration and parents living at the same address; or joint registration and parents living at different addresses/sole registration. Ethnicity was grouped into eight categories: White British; White Other; Indian; Pakistani; Bangladeshi; Black African; Black Caribbean; or Mixed/Multiple/Other. These ethnic groups were selected to enable a more detailed analysis and to avoid merging women from ethnic groups who may report diverse childbirth experiences. We included a Mixed/Multiple/Other group for women who did not identify with any of the ethnic groups specified above, but the results for this group were not interpreted in the analysis due to high heterogeneity. Country of birth was grouped into born in the UK or born outside of the UK. Those women who were born outside of the UK were further categorised according to how recently they migrated: migrated to the UK more than four years previously (before 2017) or four or fewer years previously (2017 or later). This cut-off was selected to explore experiences for women who had migrated to the UK relatively recently (women who would have been in the UK for approximately 1.5–3 years prior to becoming pregnant in 2019), and also as a pragmatic choice informed by the distribution of time since arrival while allowing for adequate sample sizes in each group for analysis. Socioeconomic deprivation was grouped according to quintiles on the IMD: 1 (least advantaged) to 5 (most advantaged). Education was grouped into three categories: completed full-time education aged 16 years or younger; aged 17 or 18 years; or aged 19 years or older.

Birth-related factors

Four birth-related factors were adjusted for in the analysis: parity (primiparous or multiparous), type of birth (single or multiple), mode of birth (vaginal, instrumental, planned Caesarean section, or unplanned Caesarean section), and gestation at birth (pre-term or term).

Childbirth experience outcomes

Women were asked about their experiences during labour and childbirth. The analysis included two composite outcomes and two single outcomes:

1. Perceived care from staff during labour and birth (composite outcome). This outcome is based on the question “Thinking about the care you received from staff during your labour and birth, how did the staff behave towards you?” Respondents are asked to consider five different aspects of staff behaviour (listed below) and each of these aspects has three response options: not at all (0), sometimes (1), or always (2). Higher scores indicate better perceived care and, for the analysis, the outcome was grouped into two categories using a cut-off point of 5/6: poorer perceived care (0–5) and better perceived care (6–10). This cut-off was selected because a score of 6 or more indicates that at least one aspect of staff behaviour had always been experienced.

2. Birth satisfaction on the Birth Satisfaction Scale - Revised Indicator (BSS-RI)¹⁹ (composite outcome). The BSS-RI is a validated scale based on six questions (listed below), each of which has three response options: agree, agree to some extent, disagree (scored 0–2 with reverse scoring on some items). Higher scores indicate greater satisfaction with care and, for the analysis, the outcome was grouped into two categories using the median score as the cut-off point: lower satisfaction (0–7) and higher satisfaction (8–12). The median has previously been used as a cut-off on the BSS-R²⁰.

3. Childbirth experience compared to expectations (single outcome). This outcome is based on the question “Overall, how do you feel your labour and birth went?” which has three response options: worse than you expected; more or less as you expected; better than you expected. For the analysis, the outcome was grouped into two categories: worse than expected or as expected/better than expected, as has been used previously²¹.

4. Overall satisfaction with care during labour and birth (single outcome). This outcome is based on the question “Overall, how satisfied or dissatisfied were you with the maternity care you received during your labour and birth?” which has five response options: very satisfied; satisfied; neither satisfied nor dissatisfied; dissatisfied; or very dissatisfied. For the analysis, the outcome was grouped into two categories: very dissatisfied/dissatisfied or other, as has been used previously²².

Analysis

For each of the childbirth outcomes, proportions and 95% confidence intervals (CI) were estimated for women from different sociodemographic groups who reported: 1) poorer perceived care during labour and birth (scoring ≤5 across the five aspects of staff behaviour); 2) lower satisfaction with care during labour and birth (scoring below the median score of 8 on the BSS-RI); 3) the birth experience was worse than expected; and 4) dissatisfaction with overall care during labour and birth. These proportions and 95% CIs are presented using a series of bar charts.

As most of the outcomes were common, univariable modified Poisson regression was used to estimate risk ratios for the association between each of the sociodemographic factors and each of the childbirth experience outcomes²³. Primarily, we wanted to explore how each of the sociodemographic factors contributed to the variation in childbirth experiences, therefore we used multivariable modified Poisson regression to mutually adjust for the sociodemographic factors (model 1). We also wanted to take account of birth-related factors that are known to be associated with childbirth experience and sociodemographic factors, and so we also adjusted for parity, type of birth, mode of birth, and gestation at birth in the multivariable model (model 2).

All analyses were conducted in Stata version 17.6²⁴ using survey-weighted commands to allow for non-response weights. Less than 1.5% of data were missing for each of the sociodemographic factors and between 0.9% and 3.9% of data were missing for the childbirth experience outcomes. As the amount of item non-response was low in comparison with unit non-response, a complete case analysis with survey weights was employed.

Characteristics of respondents

Weighted summary data describing the characteristics of the overall sample of women who responded to the 2020 national maternity survey (N=4,611) are presented in Table 1. The majority of women were aged between 25 and 34 years and the median age was 32 years (IQR=29-36 years). Over half of women registered the birth of their baby as a married couple (51.7%) and a further third (32.7%) registered the birth as both parents living at the same address. Two-thirds of women identified as White British (67.1%). Over-two-thirds of women were born in the UK (70.1%), 23.1% were born outside the UK and migrated to the UK before 2017, and 6.8% were born outside the UK and migrated to the UK in 2017 or later. The women who were born outside of the UK came from many parts of the world, principally Poland (7.3%), India (5.8%), Pakistan (5.5%), Nigeria (5.3%), Romania (4.6%) and Germany (4.5%). Just under half of the respondents were living in areas in the two least advantaged quintiles on the IMD (47.5%). Over half of women left full-time education when they were aged 19 years or older (57.7%).

Table 1 also presents the population-level distribution of these sociodemographic factors (where available), based on the overall sample of women from ONS (respondents and non-respondents) or on data from women aged 15–49 years in the 2021 census²⁵. The weighted distributions of characteristics for survey respondents are similar to the distributions in the population. However, it is noteworthy that, after applying the survey weights, women of White Other and Black African ethnicity are slightly over-represented in our survey whereas women of Indian, Pakistani and Bangladeshi ethnicity are slightly under-represented.

Childbirth experience outcomes

The majority of women reported positive experiences of childbirth and satisfaction with their care; the median birth satisfaction score on the BSS-RI was 8 (out of 12), IQR=6-10. However, 8.3% (95%CI=7.4 - 9.3) of women perceived poorer care from staff during labour and birth (score of ≤5 across the five aspects of staff behaviour), 28.3% (95%CI=26.9 - 29.9) of women experienced childbirth as worse than they had expected, and 9.1% (95%CI=8.2 - 10.1) of women reported being dissatisfied with their overall care during labour and birth. Figure 1–Figure 4 show the childbirth experience outcomes for women according to age group, registration status, ethnic group, country of birth and recency of migration, IMD quintile, and age at completion of full-time education. There is some variation in the outcomes according to the factors explored, particularly for the different ethnic groups, although there is a large amount of overlap in the confidence intervals.

Figure 1. Perceived care during labour and birth for women with different sociodemographic characteristics.

Figure 2. Birth satisfaction on the BSS-RI for women with different sociodemographic characteristics.

Figure 3. Birth experience compared to expectations for women with different sociodemographic characteristics.

Figure 4. Satisfaction with care during labour and birth for women with different sociodemographic characteristics.

Table 2 and Table 3 show the unadjusted relative risk (uRR) for the association between the sociodemographic factors and each of the composite (Table 2) and single (Table 3) childbirth experience outcomes. Table 2 and Table 3 also show the adjusted relative risk (aRR) with 95% CIs for the association between the sociodemographic factors and each of the childbirth experience outcomes after mutually adjusting for all other sociodemographic characteristics (model 1) and with further adjustment for birth-related factors (model 2). The uRR and aRR for the association between the birth-related factors and the childbirth experience outcomes are also shown.

Perceived care from staff during labour and birth

After mutually adjusting for all sociodemographic factors, there was evidence of association between registration status and ethnicity and perceived care from staff during labour and birth (Table 2, model 1). The associations remained after further adjusting for the birth-related factors (Table 2, model 2). Women who were not living with their baby’s other parent (aRR=1.99, 95%CI=1.30 - 3.04) were more likely to report poorer care from staff, compared to married women. Women of Pakistani ethnicity (aRR=2.09, 95%CI=1.74 - 3.16) were also more likely to report poorer care from staff than women of White British ethnicity.

Two of the birth-related factors were associated with perceived care from staff during labour and birth. Women who had an unplanned caesarean section (aRR=1.83, 95%CI=1.34 - 2.49) were more likely to report poorer care from staff, compared to women who gave birth vaginally, whereas women who had a multiple birth (aRR=0.21, 95%CI=0.06 - 0.74) were less likely to report poorer care from staff, compared to women who had a single birth.

Birth Satisfaction (BSS-RI)

After mutually adjusting for all sociodemographic factors, there was evidence of association between ethnicity, socioeconomic deprivation, and education and birth satisfaction on the BSS-RI (Table 2, model 1). The associations between ethnicity and socioeconomic deprivation and birth satisfaction remained after further adjusting for the birth-related factors (Table 2, model 2). Women of Pakistani (aRR=1.54, 95%CI=1.25 - 1.91), Bangladeshi (aRR=1.61, 95%CI=1.19 - 2.17) and Black African (aRR=1.41, 95%CI=1.08 - 1.85) ethnicity were more likely to report lower birth satisfaction, compared to women of White British ethnicity. Women who lived in less socioeconomically advantaged areas (within the second least advantaged quintile on the IMD) (aRR=1.14, 95%CI=1.02 - 1.28) were more likely to report lower birth satisfaction, compared to women who lived in the most socioeconomically advantaged areas. The association between education and birth satisfaction was not statistically significant after adjusting for the birth-related factors.

Three of the birth-related factors were associated with birth satisfaction. Women who had an instrumental birth (aRR=1.78, 95%CI=1.60 - 1.99) or unplanned caesarean section (aRR=2.03, 95%CI=1.84 - 2.24) were more likely to report lower birth satisfaction, compared to women who gave birth vaginally, as were women who gave birth to a pre-term baby (aRR=1.23, 95%CI=1.07 - 1.42), compared to women who gave birth at full-term.

Birth experience compared to expectations

There was evidence of association between ethnicity and country of birth/recency of migration and how women experienced the birth compared to their expectations (Table 3). Women of Indian ethnicity (aRR=0.70, 95%CI=0.49 - 1.00) and women who migrated to the UK in 2017 or later (aRR=0.66, 95%CI=0.45 - 0.98) were less likely to report that the birth was worse than they had expected, compared to women of White British ethnicity and women born in the UK, respectively and this association was statistically significant in the fully adjusted model 2.

Two of the birth-related factors were associated with how women experienced the birth compared to their expectations. Primiparous women (aRR=1.28, 95%CI=1.13 - 1.44) were more likely than multiparous women to report that the birth was worse than they had expected, as were women who had an instrumental birth (aRR=2.77, 95%CI=2.41 - 3.18) or an unplanned caesarean section (aRR=3.88, 95%CI=3.43 - 4.38), compared to women who had a vaginal birth.

Overall satisfaction with care during labour and birth

After mutually adjusting for all sociodemographic factors, there was evidence of association between ethnicity and education and overall satisfaction with care during labour and birth (Table 3, model 1). The associations remained after further adjusting for the birth-related factors (Table 3, model 2). Women of Pakistani (aRR=1.83, 95%CI=1.04 - 3.20) and Bangladeshi (aRR=2.74, 95%CI=1.29 - 5.83) ethnicity were more likely to report overall dissatisfaction with care compared to women of White British ethnicity. Women who completed their education earlier (aged 17 or 18 years) (aRR=0.74, 95%CI=0.57 - 0.98) were less likely to report overall dissatisfaction with care, compared to women who completed their education aged 19 year or older. There was also evidence of association between registration status and overall satisfaction with care during labour and birth. Women who were not living with their baby’s other parent (aRR=1.52, 95%CI=1.00 - 2.33) were more likely to report overall dissatisfaction with care, compared to married women, and this association was statistically significant in the fully adjusted model 2.

Two of the birth-related factors were associated with overall satisfaction with care during labour and birth. Primiparous women (aRR=1.40, 95%CI=1.08 - 1.82) and women who had an instrumental birth (aRR=1.42, 95%CI=1.05 - 1.91) or an unplanned caesarean section (aRR=2.10, 95%CI=1.58 - 2.79) were more likely to report overall dissatisfaction with care, compared to multiparous women and women who had a vaginal birth, respectively.

Key findings

The current study found that the majority of women reported positive experiences of childbirth and satisfaction with their maternity care; only 9% of women were dissatisfied with their overall care. For a significant minority of women, however, childbirth experiences were unsatisfactory. We found that certain sociodemographic characteristics may contribute to women’s overall experiences of childbirth. After adjusting for key birth-related factors, women of Pakistani and Bangladeshi ethnicity reported significantly poorer childbirth experiences and lower satisfaction with their care compared with women of White British ethnicity. Women of Bangladeshi origin were more than twice as likely (aRR=2.74) as women of White British ethnicity to be dissatisfied with their overall care during labour and birth, and women of Pakistani origin were twice as likely (aRR=2.09) to report poorer perceived care from staff. We also found that women of Black African ethnicity were more likely to report lower birth satisfaction, compared to women of White British ethnicity, although the findings for Black African women were inconsistent across the childbirth experience outcomes. Finally, women of Indian ethnicity tended to report more positive childbirth experiences, compared to women of White British ethnicity, but the differences did not reach statistical significance for the majority of outcomes.

The findings also show that women who were not cohabiting with their baby’s other parent reported significantly poorer perceived care from staff and lower overall satisfaction with their care during labour and birth. There was no consistent evidence of association (across at least two outcomes) between maternal age, country of birth/recency of migration, socioeconomic deprivation, or education and the childbirth experience outcomes.

In the current study, after mutually adjusting the results for all sociodemographic factors, the analysis was further adjusted for parity, type and mode of birth, and gestational age at birth as these factors are known to impact experiences of childbirth. Associations between these birth-related factors and childbirth experiences show that first-time mothers and women who had an instrumental birth or unplanned caesarean section reported consistently poorer childbirth experiences and lower satisfaction with their care. Women who had an unplanned caesarean section were almost four times more likely (aRR=3.88) to report that their childbirth experience was worse than expected and twice as likely (aRR=2.10) to be dissatisfied with their overall care during labour and birth, compared with women who had a vaginal birth. There was no consistent evidence (across at least two outcomes) to suggest that women who gave birth to a pre-term baby were more likely to report poorer childbirth experiences.

The current findings suggest that, along with certain sociodemographic characteristics, factors relating specifically to the birth are important contributors to women’s childbirth experiences.

Comparison to existing literature

Previous studies have shown that women from minority ethnic backgrounds may be less likely to have positive childbirth experiences. A nationwide maternity survey commissioned by the Care Quality Commission (CQC) in England in 2010 found that women from minority ethnic groups were less likely to feel spoken to in a way they could understand, to be treated with kindness, to be sufficiently involved in decisions, and to have confidence and trust in the staff throughout their maternity care⁶. In a qualitative study of women in England, support, communication and care needs were less likely to be met for women from minority ethnic backgrounds²⁶. A more recent survey of over 1,300 women in the UK who identified as Black or Black Mixed ethnicity found that 42% of respondents felt the standard of care they received during childbirth was poor or very poor²⁷, although there was no comparison of the findings to other ethnic groups. In the current study, a higher proportion of women of Black Caribbean ethnicity reported dissatisfaction with their overall care, compared to women of White British ethnicity (13% versus 9%, aRR=1.47), but the difference was not statistically significant, perhaps due to the relatively small number of women of Black Caribbean ethnicity in the study.

Our results are broadly in line with these previous findings, suggesting that ethnic inequalities in maternity care experiences persist, and with women from Pakistani and Bangladeshi backgrounds reporting particularly poor experiences. It is notable, however, that more recent maternity surveys by the CQC in England have not shown poorer maternity care experiences for minority ethnic groups²⁸, although outcomes are analysed by broader ethnic categories which may mask differences, such as those found for Indian, Pakistani and Bangladeshi women in the current study.

The evidence for the effect of socioeconomic factors on childbirth experience is mixed. An analysis based on data from the national maternity survey in England in 2010 found that, with decreasing socio-economic position, women were more likely to report that they were not treated respectfully or spoken to in a way they could understand by doctors and midwives²⁹. A survey in the United States also found that women with social, economic or health challenges experienced mistreatment during birth more frequently³⁰. In contrast, the most recent CQC maternity survey in England did not report variation in maternity care experiences for women with different levels of socioeconomic deprivation²⁸. Similarly, evidence from the current study on the association between socioeconomic deprivation and childbirth experiences is inconsistent.

The relationship status of parents, including whether the baby is registered in both parents’ names and whether parents are co-habiting or separated, seems to influence childbirth experiences. A large Norwegian study found that the odds of a negative childbirth experience increased for women who were not married or cohabiting⁸. This is consistent with our current study which found that women who were not co-habiting or who were sole parents had poorer childbirth experiences. We found no clear evidence of association between maternal age, education, country of birth or recency of migration and childbirth experience in the current study, which is contrary to some previous studies⁶. A systematic review of factors affecting childbirth experiences, for instance, found that age and parity were important³¹, yet a recent study of Brazilian women found no association between sociodemographic factors and level of satisfaction with the childbirth experience³².

The existing literature suggests that there is a complex array of factors that intersect to determine women’s childbirth experiences and diverse dimensions of satisfaction³³. As well as sociodemographic characteristics, several other factors have been shown to influence experiences of care. Factors relating to antenatal care, the physical birthing environment, and breastfeeding initiation have been found to be important³², as well as obstetric factors such as unexpected medical problems, low Apgar score, and transfer of the baby to intensive care^34–36. A recent review identified psychosocial factors including fear, self-efficacy, control, expectations, preparation, and support as being associated with childbirth experience³⁰. Additional factors identified include respect, privacy, inclusion in decision-making, feeling safe and nurtured, and fulfilment of personal and sociocultural expectations of childbirth^10,37.

Strengths and limitations

To our knowledge, this is one of the first studies to explore inequalities in childbirth experiences for women who gave birth in England during the pandemic. The findings are based on a large population-based sample and survey weights were employed to mitigate the effects of non-response bias and to increase the representativeness and generalisability of the findings. The survey weights only accounted for key sociodemographic factors, however, and therefore do not eliminate all sources of non-response bias from the findings. The size of the study allowed more detailed analysis of ethnic groups, rather than collapsing smaller groups into the broader categories that tend to be reported in other studies (Asian, Black, Mixed, White, and Other). Such broad merged categories can mask important differences in outcomes and experiences, such as the poorer experiences reported by Pakistani and Bangladeshi women and better experiences reported by Indian women in our study. However, using narrower sociodemographic subcategories with smaller numbers of women in each group led to wider confidence intervals around some effect size estimates. When exploring inequalities in childbirth experiences, previous studies have tended to focus primarily on ethnicity and/or socioeconomic deprivation, without considering other important factors. The current study includes multiple sociodemographic and birth-related factors and explores the interaction between these to provide further insight into the key determinants of childbirth experiences. However, we were unable to fully explore the intersection between ethnicity and socioeconomic deprivation or ethnicity and country of birth because the numbers of women in some of the groups were too small.

Implications

It is important to highlight that the majority of women who took part in the current study reported positive experiences of childbirth and satisfaction with their care. Yet, there is a clear finding that a significant minority of women have poor experiences and our findings provide evidence for sociodemographic and birth-related factors in determining differences in childbirth experiences. Women from minority ethnic backgrounds who are first-time mothers and have an instrumental delivery or unplanned caesarean section may be particularly vulnerable to poorer childbirth experiences. A recent qualitative synthesis highlighted that minority ethnic women report positive birth experiences when they are in receipt of kind, respectful and woman-centred midwifery care, but that this tends to be the exception³⁸. Therefore, in clinical practice, greater attention to basic principles of care, such as kindness and respect during labour and birth could provide more positive childbirth experiences for all women³². For women from minority ethnic backgrounds, healthcare providers should receive training in cultural sensitivity and implicit bias mitigation to ensure equitable and satisfying birth experiences for women from all backgrounds³⁹.

Women and families should expect basic care principles to be met and must be empowered to advocate for better birthing experiences. It is notable that over a quarter of all women in the current study indicated that their experience of childbirth was worse than they had expected, which suggests a key role for education and preparation during pregnancy so that women are equipped with realistic expectations about childbirth. The role of expectations is important in how childbirth is experienced and there is evidence suggesting multicultural differences in expectations for childbirth⁴⁰, which in turn can impact experience and satisfaction.

The Covid-19 pandemic caused disruption to maternity services and the restrictive measures taken had the capacity to obstruct provision of optimal maternity care around childbirth⁴¹. Comparing the findings in the current study with findings from previous national maternity surveys in England suggests that childbirth experiences were generally worse in 2020 than before the pandemic¹⁸. A qualitative study exploring the maternity care experiences of minority ethnic women who gave birth during the pandemic found that the systemic inadequacies in maternity care provision which have been identified for minority ethnic women were exacerbated by the health service modifications resulting from the pandemic⁹. It is therefore important that policy makers and health professionals are aware of the heightened risk to women from ethnic minority backgrounds during future pandemics or other times of crisis.

Ethics approval and consent to participate

The survey was approved by the North West - Greater Manchester East NRES Committee on 22^nd October 2020 (REC reference: 20/NW/0426). The NRES Committee approved the study without the requirement for informed consent. Return of partially or fully completed questionnaires was considered to imply agreement to participate in the study and consent to use the data. The participant information sheet explained that participation was voluntary and potential participants could contact the study helpline on a free telephone number should they wish to discuss any aspect of the study in more detail before deciding whether to take part.